Ansys Photonic Inverse Design Workshops

Monday, 18 October – 6:00pm-6:30pm EDT

Eli Yablonovitch, Electrical Engineering and Computer Sciences Department, University of California, Berkeley, USA

Title: Electromagnetic Design by Inverse Maxwell Solvers

Abstract: How do we co-optimize millions of geometrical parameters which define an optimal electromagnetic design? Fortunately, elegant mathematical methods have emerged that require only two solutions of Maxwell’s Equations in standard solvers to optimize a complex geometry. The mathematical approach is called the “Adjoint Method”, and it relies on the chain rule of calculus.

Monday, 18 October – 6:30pm-8:30pm EDT and Thursday, 21 October 12:00pm-2:00pm EDT

Jens Niegemann, Ansys, Canada
Taylor Robertson, Ansys, Canada

Photonic inverse design uses advanced computational methods to find optimal photonic device geometries. Designers can generate optimized devices that can be manufactured with standard photolithography techniques. Ansys’ Inverse Design workflow combines Ansys’ gold-standard Lumerical FDTD simulator with an open-source implementation of the continuous adjoint method, lumopt.

Robust, broadband, small footprint y-splitter

Join our workshop where we will perform inverse design using the adjoint optimization method to design a robust, small footprint y-splitter. Learn how to leverage a workflow combining Lumerical FDTD simulator and lumopt to design a y-splitter that can be easily manufactured and measured. The workshop will also discuss the use of open-source layout tool KLayout.

H field of an optimized design

To sign up and prepare, please:

  1. Register for the conference.
  2. Get a 30-day trial of Lumerical FDTD. Be sure to first create an Ansys account to login to
  3. Ensure you download and install the latest version of FDTD Solutions from
    • Verify that your license is correctly installed and working in advance please!
  4. Necessary if you would like to learn how to perform layout of your final design.
    • Install KLayout


  • Brief review of inverse design using the adjoint optimization method
  • Make some design choices
    • Bandwidth, number of parameters, spline boundary conditions, lithography variations
    • Advanced users may want to change parameterization entirely
  • View final S-parameter performance
    • Run python script to create full gds layout in KLayout, ready to go for cutback measurements of IL

View of the final design geometry in FDTD

View of the layout in KLayout prepared for cutback measurements

Zoomed in view in KLayout of the setup for cutback measurements